Download S1 Text

S1 Text: Supporting Materials and Methods
Embryos genotype and collection
For apical cell surface movies, wild-type embryos were w; ubi-DE-cad-GFP,
twist mutant embryos were w; twist[1], ubi-DE-cad-GFP and Kruppel mutants were
w; Kr[1], ubi-DE-cad-GFP. To obtain embryos double mutant for Kruppel and torsolike (a maternal effect mutation), w; Kr[1], ubi-DE-cad-GFP; tsl[4]/Df[3R]ED6076
females were crossed with w; Kr[1], ubi-DE-cad-GFP males. For whole embryo
imaging, genotypes of wild-type embryos were w; resille-GFP; spider GFP and twist
mutant embryos were w; twist[1]; spider-GFP. The genotype of acellular embryos
was Df(2L)dpp[s7-dp35] 21F1–3;22F1–2 (halo) Df(2L)Exel6016(slam) P{SUPorP}CG42748KG09309(CG34137)/CyOsqh–GFP [1]. Embryos were collected on grape
juice plates from flies raised at 25°C, and dechorionated in bleach prior to imaging.
Staging is according to [2]. Embryo survival and mutant phenotype was checked
systematically by allowing the imaged embryos to develop at 25°C in a humid
chamber until the end of embryogenesis. We checked that wild-type embryos hatched
as live larvae. All above mutants die at the end of embryogenesis and their phenotype
was checked by mounting cuticles in Hoyer's medium, except for acellular embryos,
where homozygous mutants were identified by the presence of the halo phenotype.
Apical cell imaging
Embryos were mounted ventrally as previously in Voltalef oil [3] and imaged
with a x40/1.3NA oil immersion objective lens on a upright Nikon E1000 microscope
coupled to Yokogawa CSU10 spinning disc confocal scanner. Illumination was with a
Spectral Applied Research LMM5 laser module (491nm excitation). Images were
captured with a Hamamatsu ImagEM EM-CCD camera driven by Volocity software
(Perkin-Elmer). For all genotypes except acellular embryos, confocal stacks of 1525µm (images separated by 1μm in z) captured the apices of the cells (some room was
left above the embryo when the stack was set up to allow for minor movement of the
embryo). For acellular embryos, stacks of 20-65µm depth were used to be able to
image the pole cells as well as the surface. Stacks were acquired every 30 seconds for
1
1-1.5 hours. Movies were recorded at 20.5
1°C measured with a high-resolution
thermometer (Checktemp1). Homozygous mutants are identifiable during imaging
based on their morphogenetic phenotype, but the phenotypes of embryos (and their
viability) were also further checked by letting the embryos develop until the end of
embryogenesis or using the halo mutation (see above). When imaging acellular
embryos, acquisition was started at the onset of gastrulation movements.
Tracking of apical cell contours
The confocal z-stacks were filtered to reduce noise (median and highpass or
tophat) and segmented in ‘o’Tracks’ as described previously [3, 4]. Automatic
tracking identified the majority of well imaged cells. Occasional misidentification was
corrected manually in some movies by deleting mistracked cell membranes and by
adding unidentified ones. Filters were also used to remove mistracked cells based on
their absolute size or their change in size (very small or large cells were excluded, as
were those that changed drastically in size from one frame to the next), their speed of
movement from one frame to the next (in order to remove cells that are incorrectly
linked in time), and the number of frames for which each cell exists. Threshold
values used for filtering were determined for each movie, carefully choosing those
values which excluded the majority of mistracked cells, but preserved as many well
tracked cells as possible. Incomplete cells at the edge of the embryo were also
removed prior to analysis.
Analysis of apical cell deformation
Analysed cells. We first selected carefully the cell populations to be analysed. In
all movies, mesoderm and mesectodermal cells were excluded from analyses based on
their dorso-ventral coordinates, taking into account the tissue translation during
extension (using so-called comoving DV coordinates). In addition for anterior
movies, cells of the head (if visible) and anterior 60µm of the trunk were excluded
based on their absolute position in embryo (static co-ordinates). Figs 1D, D’; 2A, B;
3C, D; and 5E, F show movie frame examples of the cell populations analysed after
exclusion of unwanted cells, for anterior and posterior movies (see also movies S1,
S2, S5, S6).
2
Cell shape strain rates. To measure cell shape change, a best-fit ellipse is
calculated for each cell shape at each time point. The change in shape of the cell
ellipse over time is measured as orthogonal strain rates in units of proportional rates
(pp/min) [4] (see Fig 1B). Cell shape strain rates are then projected onto embryonic
axes, AP or DV, and are called “AP or DV cell length change” in Figs and shortened
to “AP or DV cell elongation” in the main text. Note that most of the data shown in
this paper is AP cell length change, which when positive contributes to tissue
extension along the AP axis of the embryo. Proportional rate of cell area change is
calculated as the mean of the AP and DV cell length change. All colour-coded scales
are 0-0.06 pp/min (the highest value of which indicate a 6% increase in cell length or
area per minute). Strain rates calculated for each cell can be shown in movie frames
as a colour coded dot in the center of a given cell (see for example movies S1, S2, S5,
S6).
Movie synchronization. The number of embryos analysed for each experiment
was: for anterior views, 5 for wt and 5 for twi [3]; for posterior views, 4 for wt, 3 for
twi , 3 for Kr and 3 for Kr; tsl. To be able to average the data between embryos of the
same genotype and to compare different genotypes, we synchronized the movies. The
strategy for synchronization was the same as in [3], where we used the onset of
germband extension as zero for all the movies. This was determined quantitatively by
measuring total tissue strain rate in the direction of extension (this is a different
measure from cell shape strain rate) [3,4]. Because tissue strain rates fluctuate around
the onset of germ-band extension, we used a given threshold of tissue strain rate (in
the AP axis) to synchronize the movies. The threshold was the same within a given
genotype, but different between genotypes because their tissue strain rates differ. For
anterior movies, the synchronization threshold is 0.01 pp/min for wild-type and 0.005
pp/min for twist- embryos [3]. For posterior movies, we used a synchronization
threshold of 0.03 pp/min for wild type and Kruppel, 0.02 pp/min for twist and 0.01
pp/min for Kruppel; torsolike. Once all movies of the same genotype were
synchronized, we calculated the average tissue extension strain rate over time for each
genotype. Synchronisation between all movies was checked to be reasonable by
comparing timings of key events such as mesoderm invagination and mesectodermal
cell division. In the case of Kruppel and Kruppel; torsolike, the timings of these
3
events were used to adjust the intergenotype synchronisation. This was necessary for
Kruppel because for one movie (krCL070613), relatively few ectodermal cells are
visible in the field of view in the first 5 minutes of germband extension, which
affected the accuracy of the average tissue extension strain rate for this genotype at
this time. It was necessary for Kruppel; torsolike because tracking around the start of
germband extension was worse than normal for two of the movies (090713 and
100713), because of apparently lower expression of ubi-DE-cad-GFP in this
genotype, which affected the accuracy of the average tissue extension strain rate and
synchronization using the tissue threshold method.
Data summaries. Once all the movies were synchronized, we summarized data
spatially and/or temporally. The contributions of cells to all strain rate summaries
below are area-weighted. Graphs of cell shape strain rate plotted against time
summarise data from all cells included in the analysis. Graphs of cell shape strain
rate plotted against the AP axis show data for a specified timepoint or time period and
summarise all data in DV. Spatial-temporal maps (contour plots) also summarise data
in DV whilst plotting it against the orthogonal AP axis and time or summarise data in
time whilst plotting it against the AP and DV axes. Distances along embryonic axes
are given in µm from the ventral midline for the DV axis, µm from the cephalic
furrow for the AP axis in anterior movies, and µm from the posterior end of the field
of view for the AP axis in posterior movies. For the latter, because we do not image
the whole depth of the embryo, this does not correspond to the very posterior tip of
the embryo and the position of this landmark will thus depend upon the curvature of
the embryo, which might be different between genotypes (see Fig 1A’ and
discussion). Note that the black lines overlaid on spatiotemporal maps indicate tissue
translation.
Statistics. To test for evidence of differences between data derived from different
embryos, we used the mixed-effects model constructed for [3]. Briefly, we estimated
the P-value associated with a fixed effect of differences between genotypes, allowing
for random effects contributed by differences between embryos within a given
genotype, calculated at each time point or AP location. Ribbons are drawn for the
whole span of analysis for control embryos (coloured blue) and for test embryos
4
(coloured red). For spatial data, continuous data were binned into 100 equally-spaced
bins along the abscissa, for plotting and statistical tests. For the ribbons only, the
mean trends and ribbon width are calculated from data averaged to reduce noise: a
box average of +/- three bins along the abscissa were used. The widths of ribbons
straddling average strain rates represent a standard error calculated from the sums of
within-experiment variance and between experiment variance. To test where test
embryos were significantly different (P < 0.05) from wild type, mixed-model was
applied, with embryo as the random variable. No data averaging, other than binning,
was used. The regions where P < 0.05 are depicted with a grey-shaded box.
Whole embryo imaging
Embryos were washed in PBS-Tween and mounted vertically in 1.5% low
melting point agarose (Sigma) containing 0.5µm ‘yellow’ estopor microspheres
(Merck) at a concentration of 1/1000 in a cylinder using glass capillaries (Brand).
The capillary was attached to a rotation motor and immersed in a chamber filled with
PBS, where it was imaged using mSPIM at a room temperature of 28-30°C. Images
were collected every 3 µm along the z-axis, from the surface of the embryo to just
past its centre using a 20x/0.5NA water dipping objective lens (Leica) and an EMCCD camera (Andor). The illumination arms were composed of Coherent Sapphire
LP lasers (100 mW, 488nm), 1-kHz resonant mirrors, cylindrical lenses and two Zeiss
10×/0.2NA air illumination objectives. The embryo was illuminated perpendicular to
the angle of acquisition; each image was taken first with the embryo illuminated from
one side and then the other. The embryo was imaged from four angles, separated by a
90° rotation, so that all the cells of the embryo were imaged. The entire embryo was
imaged every 30 seconds for 60 minutes. The data from the four views were
reconstructed post-acquisition into a single image stack, using bead-based registration
and content based fusion (Fiji plugin) [5]. The imaging process was confirmed to
have no adverse effect on development (see above).
Temporal mapping in whole embryo movies
5
Reconstructed movies of three wild-type and three twist mutants were viewed in
4D in custom software (Browser and Tracer) written in Interactive Data Language
(IDL, Exelis) [6] and morphogenetic movements were identified by eye. The
beginning of germband extension defined as the first timepoint with detectable
posteriorward displacement of ventral cells (see Fig 4C, C’). The beginning of
mitoses in the head was defined as the timepoint in which the first cytokinesis event
in the head was observed (see Fig 4D-D”). The beginning of posterior endoderm
invagination was defined as the first timepoint in which the apices of the posterior
endoderm cells shrank detectably, and then continued to shrink in subsequent frames
(see Fig 4E, E’). The beginning of mesodermal tube sealing was defined as the
timepoint when the right and left sides of the tissue first met to begin forming the
internal mesodermal tube (see Fig 4E, E’). The beginning of dorsal fold formation
was defined as the first timepoint at which detectable buckling (ie basalward
movement of cell apices) could be seen on the dorsal side of the embryo (see Fig 4G,
G’). The beginning of dorsal contraction was defined as the first timepoint at which
dorsal anterior and dorsal posterior cells could be detected moving towards each other
(see Fig 4H, H’).
Antibody stainings of acellular embryos
We followed standard methods as in [7] for fixing and staining
acellular embryos, using the primary antibodies anti-DE-Cadherin (1/50, DCAD2,
developed by T. Uemura, Developmental Studies Hybridoma Bank) and anti-Sqh1P
(1/100, a gift from R. Ward,[8]).
Particle Image Velocimetry of Myosin II flows in acellular embryos
Acellular embryos expressing sqh-GFP to label Myosin II were selected during
cellularisation based on their halo phenotype, mounted laterally and imaged as
described above. A plugin in FIJI was used to perform Particle Image Velocimetry
(PIV) of images of acellular embryos [9]. Stacks were processed to maximumintensity Z-projections, and background signal from outside the embryo was removed
6
manually from the images. Three iterations of PIV were run using the normalized
correlation coefficient method, with interrogation window halving each time, giving a
grid spacing of 8.7 × 8.7 µm in the plotted displacement field.
Laser ablation of acellular embryos
Mounting acellular embryos as above, laser ablation experiments were
performed using a TriM Scope II Upright 2-photon Scanning Fluorescence
Microscope controlled by Inspector Pro software (LaVision Biotec). The laser source
for the microscope was a tuneable near-infrared (NIR) laser delivering 120
femtosecond pulses with a repetition rate of 80 MHz (Insight DeepSee, SpectraPhysics). The laser was tuned to 927nm, with an average power of 1.7 W. The
maximum laser power allowed to reach the sample was set to 190 mW and an
Electro-Optical Modulator (EOM) was used to allow microsecond switching between
imaging and treatment laser powers. The laser light was focused by a 25x, 1.05
Numerical Aperture (NA) water immersion objective lens with a 2mm working
distance (XLPLN25XWMP2, Olympus). Given the non-linearity of the 2-photon
absorption process an NA of 1.05 allows incident laser light of 927nm wavelength to
be focused to a spot with FWHM of 340nm laterally and 1.2µm axially, although the
spatial extent of treatment at high powers will likely exceed this volume. Images were
collected every 0.742ms for 20 frames before the ablation and 120 frames after the
ablation, using a GaAsP photomultiplier tube.
Targeted line ablations of 20µm length were performed on the apical Myosin II
mesh (approximately 2µm below the vitelline membrane) using a treatment power of
190 mW. Ablations were performed during image acquisition (with a dwell time of
10µsec per pixel), with the laser power switching between treatment and imaging
powers as the laser was raster scanned across the sample.
Line ablations were
oriented parallel to the dorso-ventral embryonic axis (DV ablations) or parallel to the
anterior-posterior embryonic axes (AP ablations) and were performed both near the
7
posterior tip of the embryo (posterior ablations) or near the middle of the anteriorposterior axis of the embryo (anterior ablations).
Analysis of recoil velocities from laser ablations
In order to quantify recoil, we had to automatically detect the position of the
ablation line in the images, choose a region of interest around it, detect the fluorescent
structures inside this region and infer their motion from the video sequences. The
region of interest in each movie is defined as a 20 pixel area on each side of the cut.
The position of the cut is determined by thresholding the first image acquired after the
cut (timepoint 21, +0.742 seconds) with a threshold estimated from a Gaussian model
of pixel intensities. The Myosin II signal appears in the image as small fluorescent
bright structures on dark background. The bright pixels are considered true signal if
the à trous wavelet coefficients are deemed significant at two consecutive levels by a
selection algorithm based on the control of the false discovery rate [10]. Only
embryos with more than 100 significant pixels of signal in the region of interest were
included in the study. We first performed two pre-processing steps: a frequency filter
to remove the stripes caused by electrical interference in the detector and a patch
based denoising step [11]. Subsequently, we computed the optical flow in order to
estimate the velocity of the structures using the Lucas-Kanade algorithm [12], on the
significant pixels of signal in the region of interest for each embryo. Square windows
of 12 pixels width were used in the computation of the optical flow. In order to use
the flow of Myosin II to measure recoil away from the cut, only the velocity
component of flow perpendicular to the cut is considered in the analysis. To take
account of translation, the velocity at time point 22 (+ 1.484 seconds) was corrected
by subtracting the average speed in the region of interest before ablation, computed
from the optical flow of the three frames before the cut (timepoints 18, 19 and 20, 2.226, -1.484 and – 0.742 seconds) to give the normalised relaxation speed. For each
side of the cut, the average normalised relaxation speed is computed and the two
averages are combined (weighted average, taking account of the number of pixels of
significant signal on each side of the cut) to give a measure of motion away from the
cut (final normalised relaxation speed shown in Fig 6E). This final velocity was
8
compared for the four conditions and a two sample t-test was used to look for
statistically significant differences between groups (Fig 6E).
1.
He B, Doubrovinski K, Polyakov O, Wieschaus E. Apical constriction drives
tissue-scale hydrodynamic flow to mediate cell elongation. Nature.
2014;508(7496):392-6. Epub 2014/03/05. doi: 10.1038/nature13070. PubMed
PMID: 24590071; PubMed Central PMCID: PMC4111109.
2.
Wieschaus E, Nusslein-Volhard C. Looking at embryos. In: Roberts DB,
editor. Drosophila, a practical approach. United States: Oxford University Press
Inc., New York; 1998.
3.
Butler LC, Blanchard GB, Kabla AJ, Lawrence NJ, Welchman DP,
Mahadevan L, et al. Cell shape changes indicate a role for extrinsic tensile forces
in Drosophila germ-band extension. Nature cell biology. 2009;11(7):859-64.
Epub 2009/06/09. doi: 10.1038/ncb1894. PubMed PMID: 19503074.
4.
Blanchard GB, Kabla AJ, Schultz NL, Butler LC, Sanson B, Gorfinkiel N, et al.
Tissue tectonics: morphogenetic strain rates, cell shape change and intercalation.
Nature methods. 2009;6(6):458-64. Epub 2009/05/05. doi:
10.1038/nmeth.1327. PubMed PMID: 19412170.
5.
Preibisch S, Saalfeld S, Schindelin J, Tomancak P. Software for bead-based
registration of selective plane illumination microscopy data. Nature methods.
2010;7(6):418-9. Epub 2010/05/29. doi: 10.1038/nmeth0610-418. PubMed
PMID: 20508634.
6.
England SJ, Blanchard GB, Mahadevan L, Adams RJ. A dynamic fate map of
the forebrain shows how vertebrate eyes form and explains two causes of
cyclopia. Development. 2006;133(23):4613-7. Epub 2006/11/03. doi:
10.1242/dev.02678. PubMed PMID: 17079266.
7.
Lye CM, Naylor HW, Sanson B. Subcellular localisations of the CPTI
collection of YFP-tagged proteins in Drosophila embryos. Development. 2014;In
press.
8.
Zhang L, Ward REt. Distinct tissue distributions and subcellular
localizations of differently phosphorylated forms of the myosin regulatory light
chain in Drosophila. Gene expression patterns : GEP. 2011;11(1-2):93-104. Epub
2010/10/06. doi: 10.1016/j.gep.2010.09.008. PubMed PMID: 20920606;
PubMed Central PMCID: PMC3025304.
9.
Tseng Q, Duchemin-Pelletier E, Deshiere A, Balland M, Guillou H, Filhol O,
et al. Spatial organization of the extracellular matrix regulates cell-cell junction
positioning. Proceedings of the National Academy of Sciences of the United
States of America. 2012;109(5):1506-11. Epub 2012/02/07. doi:
10.1073/pnas.1106377109. PubMed PMID: 22307605; PubMed Central PMCID:
PMC3277177.
10.
Muresan L, Jacak J, Klement EP, Hesse J, Schutz GJ. Microarray analysis at
single-molecule resolution. IEEE transactions on nanobioscience. 2010;9(1):518. Epub 2010/02/04. doi: 10.1109/TNB.2010.2040627. PubMed PMID:
20123580; PubMed Central PMCID: PMC2912528.
11.
Boulanger J, Kervrann C, Bouthemy P, Elbau P, Sibarita JB, Salamero J.
Patch-based nonlocal functional for denoising fluorescence microscopy image
9
sequences. IEEE transactions on medical imaging. 2010;29(2):442-54. Epub
2009/11/11. doi: 10.1109/TMI.2009.2033991. PubMed PMID: 19900849.
12.
Lucas DB, Kanade T, editors. An Iterative Image Registration Technique
with an Application to Stereo Vision. International Joint Conference on Artificial
Intelligence; 1981.
10